As for point 1, I think it would be better that the user see values that have the same units, but you should be able to do something like the following:
For point 2, investigate the class function in QV help or in the forum. You have to make a calculated dimension with the class function and put the count as an expression.
It works!!! I use class function and it works ok when I just select one "service" in order to know his response time distribution. But I've problems selecting more than one service when "relative" is activated. (and I really need it!)
My dimensions are:
=if( performance_data.value<3, class(performance_data.value, 0.5,'segons'))
And my expression is:
With this configuration:
a) selecting "service A" it has 80% of response time between 0-0,3s
b) selecting "service A and service B", service A has just 25% of response time between 0-0,3s
I suppose that Qlikview applies "relative" with all information data and not for each dimension. Isn't it? How could I solve it?
Thank ypu so much!!!
Yeah, instead of using the relative option in QlikView you would have to do something like the following to always return the porcent per each service:
count(Total <servicecheck_name> performance_data.value)
I don't know know if this will work correctly with the dimension you have defined, but I hope it points you in the right direction. If you need more help upload an example of your application.
Thank you for your answer but I'm calcuating a distribution chart, so I can't change the expression...
I need the % of occurrencies in each range.
I don't know what info could I upload, so I'll try to give you an easy example:
1 Service A
2 Service B
ID Time value
1 d1 0,45
1 d2 0,3
1 d3 0,75
2 d4 0,75
2 d5 0,15
2 d6 1,3
So, using the previous dimension ans expression definition I was trying to get that results:
* selecting just ServiceB:
Range Ocurrencies %
0-0,5 2 66%
0,5-1 1 33%
* selecting ServiceA and ServiceB:
Range Service Ocurrencies %
0-0,5 ServiceA 2 66% (and using relative I get 33%)
0-0,5 ServiceB 1 33% (and using relative I get 16%)
0,5-1 ServiceA 1 33% (and using relative I get 16%)
0,5-1 ServiceB 1 33% (and using relative I get 16%)
1-1,5 ServiceB 1 33% (and using relative I get 16%)
So if I try to make a % manually modifying the expression, I won't calculate time in each range...
Isn't it? How could I get it?
Thank you so much.
I used the formula I recommended above and it seemed to work. See the attached example.
distribution.qvw 133.2 K
Here's the picture of the table. Using the same column names and data as above, the dimensions are:
The expressions are:
count(distinct Time)/count(distinct total <name> Time)
distribution.JPG 21.6 K